Introduction
SANA-WM is an efficient open-source world model trained natively for one-minute generation. The bidirectional checkpoint released here is a 2.6B-parameter image-to-video diffusion transformer that synthesises 720p, minute-scale videos with precise 6-DoF camera control, paired with the LTX-2 sink-bidirectional Euler refiner for high-fidelity decoding.
Core Architecture Designs
Four core designs drive the architecture:
- Hybrid Linear Attention — frame-wise Gated DeltaNet combined with softmax attention every Nth block for memory-efficient long-context modelling.
- Dual-Branch Camera Control — independent main and camera branches enable precise per-frame trajectory adherence.
- Two-Stage Generation Pipeline — a long-video refiner stitched on top of Stage-1 latents improves quality and temporal consistency.
- Robust Annotation Pipeline — metric-scale 6-DoF camera poses extracted from public video corpora yield spatiotemporally consistent action supervision.
@article{zhu2026sanawm, title = {{SANA-WM}: Efficient Minute-Scale World Modeling with Hybrid Linear Diffusion Transformer}, author = {Zhu, Haoyi and Liu, Haozhe and Zhao, Yuyang and Ye, Tian and Chen, Junsong and Yu, Jincheng and He, Tong and Han, Song and Xie, Enze}, journal = {arXiv preprint arXiv:2605.15178}, year = {2026}, }
Repository Layout
| Component | Path in repo | Size |
|---|---|---|
| Sana DiT (Stage 1) | dit/sana_wm_1600m_720p.safetensors | 10 GB |
| LTX-2 VAE (diffusers) | vae/ | 2 GB |
| LTX-2 refiner (Stage 2) | refiner/refiner.safetensors | 41 GB |
| Gemma text encoder for the refiner | refiner/text_encoder/ | 46 GB |
| Inference config | config.yaml | — |
The Sana text encoder (gemma-2-2b-it) is not bundled here — it is fetched on demand from the public Hugging Face mirror.
python inference_video_scripts/inference_sana_wm.py \ --image asset/sana_wm/demo_0.png \ --prompt asset/sana_wm/demo_0.txt \ --action "w-80,jw-40,w-40,lw-60,w-100" \ --translation_speed 0.055 \ --rotation_speed_deg 1.2 \ --num_frames 321 \ --output_dir results/demo
Usage Details
Weights are fetched from this repository on first use. Pass --no_refiner to skip the LTX-2 refiner and decode Stage-1 latents with the Sana VAE instead. To run fully offline, override any of --config / --model_path / --refiner_checkpoint / --refiner_gemma_root with local paths.
The output frame size is fixed at 704 x 1280; input images are aspect-preserving resized + center-cropped to that resolution.
Inputs
| Argument | Format |
|---|---|
--image | RGB image (any PIL-readable format) — used as the first frame. |
--prompt | UTF-8 text file containing the conditioning prompt. |
--camera | NumPy .npy of shape (F, 4, 4) — per-frame camera-to-world matrices. |
--action | WASD/IJKL DSL, e.g. "w-80,jw-40,w-40,lw-60,w-100". We roll it out to a (F+1, 4, 4) trajectory. Mutually exclusive with --camera. |
--intrinsics | Optional. .npy of shape (3, 3), (F, 3, 3), or (4,). If omitted, we estimate intrinsics from --image with Pi3X and abort if the resulting FOV is outside [25°, 120°]. |
Social Feedback
Social media reports highlight the following user-observed characteristics and performance:
- Runs on less than 8 GB VRAM.
- A distilled version denoises a 60-second 720p clip in 34 seconds on one RTX 5090.
- Reported to be 36× faster than older open models.
- Trained on approximately 213K public videos for 15 days on 64 H100 GPUs.
- Compatible with ComfyUI and Diffusers plugins.
- Licensed under Apache 2.0.
- One user reported receiving only a regular video output (potential quality or correctness issue).
